Dominikus Baur is a data visualisation expert and interaction designer from Germany. He creates usable, aesthetic and responsive visualisations for desktops, tablets and smartphones. In September he will deliver a training in Utrecht on Online Data Visualisation. Dominikus: “Visualisation people – like all experts – have a tendency to see every dataset as a potential nail for their visualisation hammer”.
Graphic Hunters: What is the power of a good visualisation? In what way can a visualisation help to understand or to communicate information?
Dominikus Baur: We humans are a bit in over our heads when it comes to the current age of data. We’re not really capable of understanding and processing large datasets, since evolution has shaped us for a much simpler lifestyle (something with savannas and berries) and we’re usually stymied by even a small number of data points. Our minds are simply not well equipped for all things numerical.
But instead of giving up and handing understanding the issues of our time over to the black boxes of machine learning, we can use visualisation, our best tool for bridging the gap between our abilities and the amount of data we have.
Visualisation takes data and turns it from something numerical (which we’re bad at) into something visual (which we’re extremely good at). Basically hacking our senses, we’re transforming data into pictures and switching from the narrow channel of abstract thinking to the broadband channel of visual processing.
And this trick becomes even better when we’re combining these pictures with interaction, letting us use the computer’s capabilities to dynamically transform them into answers to our questions, one after the other.
I have read that someone has to understand a visualisation within seconds, others say visualisations should leave something to explore. What is your opinion on this?
The very best visualisations do both! Understanding what dataset is shown and what the visual elements depict is key to making a visualisation accessible and engaging the audience right away. Having some introductory text explaining what’s happening is perfectly fine, but the graphic should also be able to stand on its own. Guiding the audience into the visualisation can also happen in a series of steps, highlighting various aspects of the data. At first glance, the overarching patterns should also become visible, already leading to some insight into the underlying dataset.
Once people are interested in the data, it can be very frustrating for them if that’s all there is. It constitutes a lost opportunity if what people learned about the data in the very first moments is also the total amount of depth the visualisation provides. Even if not everybody has the time or interest to dig deeper, visualisations are also great tools for exploration. Giving people interactive means to shape their own lenses on the data and ask questions they are interested in separates good visualisations from great ones.
What is the first most important question one should ask before starting visualising the data?
The first question should be: why are we visualising this data? If you look closely, this question consists of two essential aspects. The more obvious one is: once we have our visualisation, how are people going to use it? Which questions about the data are they going to answer with this tool? What types of visualisations and interaction can we use to help our audience with their tasks?
Yet, the subtler connotation of the question is: is visualisation really the right tool for this job? Visualisation people – like all experts – have a tendency to see every dataset as a potential nail for their visualisation hammer. But oftentimes, other tools for making sense of data – like statistics – are perfectly fine and maybe even more straightforward in delivering the answers the audience
is looking for.
All in all, knowing what the audience’s context and goals will be should be the central aspect for such decisions. Intricate visualisations for somebody who’s in a hurry and only wants a one line answer can cause frustration, just like a dumbed down tool for people who want to dig into the data. Having a clear picture of your audience and their situation can make creating a successful visualisation much easier.
If you could choose one visualisation that you can put in a frame and hang in your livingroom, which visualisation would you choose and why?
While it would be obvious to answer this question with one of the classics that appear in every text book and introduction to visualisation, my personal favorites would be hard to hang in a livingroom without taking away their essence: interactivity and animation. These aspects were what first drew me to visualisation as a profession.
My favorite from the (relatively) early days of interactive visualisation is “We Feel Fine” by Jonathan Harris and Sep Kamvar (http://wefeelfine.org/). The project shows human emotions as expressed online in a group of dynamic and playful visualisations (called movements). Single sentiments are often depicted as bubbles that can be interacted with to dig down to the original posts. We Feel Fine shows visualisation at its best: it is immediately accessible and understandable, while also enabling in-depth exploration. Additionally, it proved that online visualisation could work and paved the way for all current online data journalism and web-based visualisation projects.
The only problem these days is that all the blogs We Feel Fine is based on have been abandoned a long time ago, making it a visualisation without a dataset. Maybe it’s time for a 2017 version based on Twitter data or so.
What is the most frequently asked question in your training sessions and what is your answer to this?
One of the questions I often get in introduction courses is: “When are we talking about d3.js?” This framework has made a tremendous impact on data visualisation as a whole and also its public image. In many people’s heads online data visualisation and d3.js are more or less the same.
But something as powerful as d3.js of course also takes a while to explain, so in these introduction courses my answer usually is: “Sorry”. Therefore I’m very excited about the upcoming training, where we will have two full days to discuss not only d3.js but also online data visualisation in general in much more detail!
De training Online Data Visualisation staat gepland op 28 en 29 september 2017 in Utrecht. Enthousiast geworden over deze training? Kijk voor meer informatie op de website en schrijf je in.